Research map / Video / Video generation
Generating Humanless Environment Walkthroughs from Egocentric Walking Tour Videos
Yujin Ham, Junho Kim, Vivek Boominathan, Guha Balakrishnan
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| fine-tune Casper on EgoCrowds | Casper · EgoCrowds test | PSNR (avg): 24.64 -> 26.74 | Improved |
| fine-tune Casper on EgoCrowds | Casper · EgoCrowds test | DreamSim (avg): 0.029 -> 0.022 | Improved |
| fine-tune Casper on EgoCrowds | Casper · EgoCrowds test | LPIPS (avg): 0.130 -> 0.118 | Improved |
| add shadow injection to training data (ablation) | CrowdEraser · EgoCrowds test | PSNR: 25.09 -> 26.74 | Improved |
| add shadow injection to training data (ablation) | CrowdEraser · EgoCrowds test | DreamSim: 0.028 -> 0.022 | Improved |
| train on uniform Crowd% coverage instead of 0-10% only (ablation) | CrowdEraser · EgoCrowds test | PSNR: 26.60 -> 26.74 | Improved |
| train on uniform Crowd% coverage instead of 0-10% only (ablation) | CrowdEraser · EgoCrowds test | SSIM: 0.882 -> 0.881 | No clear effect |
| use CrowdEraser vs ProPainter/DiffuEraser at high crowd | CrowdEraser · EgoCrowds test (Crowd% 40-50%) | PSNR (Crowd% 40-50): 22.27 -> 23.31 | Improved |
| use CrowdEraser vs baselines at low crowd | CrowdEraser · EgoCrowds test (Crowd% 0-10%) | LPIPS (Crowd% 0-10): 0.055 -> 0.065 | Hurt |
| use cleaned videos as input to SpatialTrackerV2 | SpatialTrackerV2 · Marrakech/Bangkok/Istanbul clips | 3D reconstruction quality | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ProPainter — ProPainter: Improving Propagation and Transformer for Video Inpainting (ICCV 2023) · better on 5 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- EgoGen: An Egocentric Synthetic Data Generator (CVPR 2024 · similarity 0.85)
- EgoVid-5M: A Large-Scale Video-Action Dataset for Egocentric Videos Generation (NeurIPS 2025 · similarity 0.85)
- Learning to Generate Diverse Pedestrian Movements from Web Videos with Noisy Labels (ICLR 2025 · similarity 0.85)
- PlayerOne: Egocentric World Simulator (NeurIPS 2025 · similarity 0.84)
- Exocentric-to-Egocentric Video Generation (NeurIPS 2024 · similarity 0.84)
- HumanVid: Demystifying Training Data for Camera-controllable Human Image Animation (NeurIPS 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
